Defect detection in thin-film photovoltaics; Towards improved efficiency and longevity

The Photovoltaic (PV) industry is seeking to increase efficiency and functional lifetime of PV modules manufactured on polymer substrates. High resolution and high speed surface inspection for the quality control of the manufacture of large area flexible PV modules are necessary to guarantee maximum quality, longer lifetime and enhanced product yield. Flexible PV films are the newest development in the renewable energy field and the latest films have efficiencies at or beyond the level of Si-based rigid PV modules. However, they are at present highly susceptible to long term environmental degradation as a result of water vapor transmission through the protective encapsulation to the active layer. To reduce the water vapor transmission rate (WVTR) the PV encapsulation includes a barrier layer of amorphous Al₂O₃ on a planarised polymer substrate. This highly conformal barrier layer is produced by atomic layer deposition (ALD). Nevertheless water vapour transmission is still facilitated by the presence of micro and nano-scale defects in these barriers which results in decreased cell efficiency and reduced longevity. The main aim of this research paper is to use surface metrology techniques including: White Light Scanning Interferometry (WLSI), Atomic Force Microscopy (AFM) and Scanning Electron Microscopy (SEM) to characterise the water vapor barrier defects which are seemed to be responsible for the water vapor permeation. A real surface texture parameter analysis allows the efficient separation of small insignificant features from significant defects. This parametric analysis is then correlated with the water vapour transmission rate as measured on typical sets of films using standard MOCON test. The paper finishes by drawing conclusions based on analysis of WVTR and defect size, density and distribution, where it is postulated that small numbers of large features have more influence on the deterioration of water vapor transmission rates than l- rge numbers of small features. This result provides the basis for developing roll-to-roll in process metrology devices for quality control.